Our molecular epidemiology studies are focused on DNA repair gene polymorphisms and on the measurement of repair capacity; to identify important risks for environmentally-associated cancers. The group of Intramural laboratories working on repair provides a strong basic science foundation of expert colleagues that facilitates our ability to translate studeis to the population level and examine the effects of polymorphisms and repair capacity on mutation and cancer risk. Research on genetic susceptibility and gene-environment interaction may identify variants of genes important to carcinogenesis and the pathways by which environmental agents damage DNA. The study of genetically susceptible subgroups may allow a more precise identification of environmental exposures that cause disease and the risk from such exposure. Finally, if important susceptibility genes are identified, it could lead to public health programs for protecting susceptible populations, and for targeted screening of groups at higher risk of disease. Bladder Cancer XRCC3. We examined the role of a common polymorphism in the XRCC3 gene (codon 241: threonine to methionine change) and bladder cancer risk. This gene plays a role in the homologous recombination pathway, that repairs double strand breaks (DSBs). We hypothesized that the codon 241 polymorphism could affect repair of smoking-associated DNA damage and could thereby affect bladder cancer risk. We genotyped 233 bladder cancer cases and 209 controls who had been frequency matched to cases on age, sex, and ethnicity. We observed a weak positive association between subjects who carried at least one copy of the codon 241 Met allele and bladder cancer (OR: 1.3; 95% CI: 0.9-1.9). Among heavy-smokers, individuals with the Met allele had about twice the risk of those without it. Previously, we observed in these subjects an association between bladder cancer risk and allelic variants of the XRCC1 gene, which is involved in the repair of base damage and single strand breaks (SSBs). Interestingly, we find evidence of interaction between these two genes on bladder cancer risk and some support for a gene-gene-smoking exposure interaction on risk. XPD. We hypothesized that an XPD codon 751 polymorphism (Lys to Gln amino acid change) could affect the repair of smoking-induced DNA damage and could be associated with bladder cancer risk. We determined the XPD codon 751 genotype for 228 bladder cancer cases and 210 controls who were frequency matched to cases by age, sex, and ethnicity. We found a slight decrease in risk for the XPD codon 751 Gln/Gln genotype (adjusted OR: 0.8; 95% CI: 0.4-1.3) compared to subjects with the Lys/Lys or Lys/Gln genotypes. The analysis with smoking showed that smokers with the Lys/Lys or Lys/Gln genotypes were twice as likely to have bladder cancer than smokers with the Gln/Gln genotype (test of interaction p= 0.03). HRAS1. To examine whether individuals with rare HRAS1 VNTR alleles are at increased risk of bladder cancer, we carried out a case-control study with 230 bladder cancer cases and 203 hospital-based controls. HRAS1 genotype may be related to the prognosis of bladder cancer because incident cases had a higher frequency of rare alleles than did prevalent cases. Prostate cancer. We proposed a hypothesis that reduced base excision repair capacity modulates the effect of diet-associated oxidative damage on prostate cancer risk. We examined whether polymorphisms in the XRCC1 gene affect prostate cancer risk. Men who were homozygous for the XRCC1 codon 399 Arg allele showed a slightly higher prostate cancer risk than those with one or two copies of the GLN allele. Across different strata of dietary intake, this risk was slightly but remarkably consistently increased. The risk was highest among men with the Arg/Arg genotype and low dietary intake of vitamin E or lycopene; whereas low intake of these antioxidants in men without this genotype hardly increased prostate cancer risk. Functional Measures of DNA Repair for Population Studies Measure of Recombinational Repair. In order to be able to measure individual differences in recombinational DNA repair we developed a novel assay based on recombination between two Green Fluorescent Protein (GFP) sequences in transiently transfected plasmid DNA. The plasmid construct contains an intact, emission-shifted, 'blue' variant of GFP (BFP), with a 300 nucleotide stretch of homology to a nonfunctional copy of GFP. In the absence of homologous recombination only BFP is present, but homologous recombination can create a functional GFP. The homologous regions in the plasmid were constructed in both the direct and the inverted orientation of transcription to detect possible differences in the recombination mechanisms involved. A panel of human tumor cell lines was chosen on the basis of genetic background and chromosome integrity and tested for homologous recombination using this assay. The panel included cell lines with varying levels of karyotypic abnormalities, isogenic cell lines with normal and mutant p53, isogenic cell lines with or without DNA mismatch repair, BRCA1 and -2 mutant cell lines, and the lymphoma cell line DT40. With this assay, the observed differences between cell lines with the lowest and highest levels of recombination were about 100-fold. Increased levels of recombination were associated with mutant p53, whereas a low level of recombination was found in the BRCA1 mutant cell line. In the cell line HT1080TG, a mutagenized derivative of HT1080 with two mutant alleles of p53, high levels of recombination were found with the direct orientation but not with the inverted orientation plasmid. No difference in recombination was detected between two isogenic cell lines that only differed in DNA mismatch repair capability. Functional Measure of DNA Repair: We have established the Single Cell Gel Electrophoresis (Comet) assay in my laboratory for measuring DNA damage and repair. ). In this assay, DNA is damage introduced through in vitro exposure of cells to different chemical carcinogens. Cells are embedded in agarose on a microscope slide, lysed, and the remaining nucleoids are subject to electrophoresis. Nucleoids from cells without DNA damage remain spherical, while those with damaged DNA have migration of the free DNA strands through the agarose. When visualized using a fluorescent microscope these cells have a comet-like shape, with the tail formed from broken DNA. The amount of DNA in the tails of individual cells can be rapidly and precisely measured using image analysis software. By allowing different lengths of repair time before cells are lysed, we can precisely measure repair kinetics. We are beginning to apply these measures to lymphocytes from cases and controls in the bladder cancer study in order to assess risk from decreased repair capacity and to assess the consequences of newly described polymorphisms in repair gene on direct functional measures of repair. Part of the power of Comet analysis coupled with automated image analysis software is that it provides multiple DNA damage outcome measurements on each of 100 individual cells from a particular treatment. Data Analysis. Motivated by the single cell gel or comet assay application, we proposed a general approach for Bayesian quantile regression of clustered data. To simplify modeling, the likelihood is approximated by a substitution likelihood, which depends on a vector of unknown quantiles. Covariate effects and heterogeneity among samples in these quantiles are modeled using a Gaussian hierarchical model, with appropriate order constraints, and a Markov chain Monte Carlo algorithm for a posterior computation.